NCT00088699 Strong evidence from family and twin studies demonstrates that major depressive disorder (MDD) is heritable, yet there has been limited progress in identifying the actual genes involved. A separate, perhaps overlapping set of genes is expected to play a role in individual variation in treatment response in MDD. By use of a large set of markers in many genes, we seek to characterize patients who differ in their response to standard antidepressant treatments. In the first years of this project, candidate gene studies implicated a few genes in treatment outcome and other genes in adverse effects. Future studies are needed to determine whether individuals who carry such genetic markers may benefit from closer monitoring or alternative treatments. We also participated in a meta-analysis of three genome-wide association studies of antidepressant outcome. Despite greater power of this combined sample to uncover association with common genetic markers, no genome-wide significant associations were found. We concluded that no common alleles of large effect on antidepressant outcome exist in these samples. Recently, we re-examined this question in the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) sample using newer analysis methods that improve upon older genotyping techniques initially used in this sample. In collaboration with Dr. Yin Yao and colleagues at NIMH, we carried out a new genome wide association study of antidepressant response in the STAR*D sample.. We identified several SNPs associated with outcome after 12 weeks of treatment with citalopram, and several different SNPs that were associated with outcome after 12 weeks. Heritability was greater for outcomes following longer treatment periods, an unexpected finding that suggests the shorter treatment durations typically used in pharmacogenomic studies may not be optimal. Ongoing work is aimed at replicating these finding in additional samples. We are also using new, high-throughput sequencing methods to test for rarer genetic variants that may exert larger effects, at least in patients with treatment resistant depression. Such variants may show larger effects, especially among patients with unusual treatment outcomes. Patients who respond to antidepressant treatment constitute a mixture of true responders, placebo responders, and those whose illness remits spontaneously. Patients who fail to respond to multiple treatments may be less heterogeneous, since placebo responders and spontaneous remitters are removed. From over 4,000 patients enrolled in the STAR*D study, we found that only 10% show treatment resistance that is not explained by non-adherence, comorbid substance use disorder, or other factors, and only 3% are highly treatment-resistant. Sequencing of the coding regions of the genome (exome) has now been completed on over 175 treatment-resistant (TRD) or typically responsive patients drawn from the STAR*D, the Univ. Michigan Depression Center, and NIMH studies of novel antidepressants such as ketamine and scopolamine. Much of the exome sequencing was carried out at the NIH Intramural Sequencing Center with funds provided by the NIH Clinical Center Genomics Opportunity (CCGO) program, which also provided exome sequence from about 200 non-psychiatric patients for comparison. Of the 350,000 high-quality genetic variants identified, about 15,000 were found to be rare, potentially-damaging variants within the protein-coding regions of genes. Some of these variants appeared in up to 6 TRD cases, although they were not detected in other sequenced patients. Nevertheless, no variant, gene, or gene set was found to be significantly associated with TRD after correction for multiple testing. In the coming year, we plan to expand the sample size, incorporate information on common genetic variants, identify copy number variants, and consider variants in non-coding regions. Taken together, these approaches may increase power and identify genes more robustly associated with resistance to anti-depressant treatment.